Vector Deviations "T"

[tdriesenga]

Can someone tell me when I take a staright vetor "X" axis edge point and then report "X" axis deviation & vector deviation "T" the deviations do not exactly match???? I am seeing up to a .01mm difference on a .9999 something approach vector.

[Matthew D. Hoedeman]

1) Are you using SNAP points? If not, then machine drift will ADD to the T value (I always use SNAP points)

Report the Y and Z deviations as well, then use TRIG:
SQRT(X DEV^2 + Y DEV^2 + Z DEV^2)
I'll bet it matches the T deviation then.

[tdriesenga]

Thanks for the suggestion but auto edge point does not support SNAP points I discovered! Any other ideas?

[jmgreen]

The square root method does not work with T as I've discovered. Attached is the BnS explanation of where T comes from. I have gone round and round about this with Tech support and I now just tell people it is the deviation from nominal. In MM4 you could actually do the square root calculation to demonstrate where the numbers come from. I think this will show you why it varies slightly.

[Matthew D. Hoedeman]

OK, if you are using a surface sample hit, you will be getting surface deviation added into the results. HOWEVER, for the dimension, you can use GAP ONLY (toggle button when creating the dimension) and it will 'remove' the surface deviation from the dimension. This means that it uses the surface sample hit to set the depth of the touch on the edge point (and you SHOULD use at least 1 surface sample touch), then 'projects' the point back to the nominal surface location, thus removing the deviations caused by an imperfect surface.

[Thomas Stewart]

I agree with Matt. Now I'll never profess to be a wiz banger as Matt or other high rankers on here are, but I've found consistantly accurate results in edge point with GAP on... although I've almost exclusively used the 'S' value when tolerancing (per PC-DMIS training instructor), I have gotten good results ever since.